Do racial profiling tactics differ by officer race? The literature has relied on a test based on hit rates in vehicle searches to answer this question. This paper instead argues for a test based on patrol officers' stop rates since it is less manipulable by officers, requires fewer assumptions, and includes all officer-citizen interactions. I compare these two methods of testing for discrimination using detailed data from Oakland, California. The hit rate test concludes that officer race does not matter on average or within neighborhoods. In contrast, the stop rate test concludes that officer race does matter within neighborhoods, and further suggests that the same officers discriminate in favor of their own race in some neighborhoods, yet against their own race in other neighborhoods. The contradictory nature of this discrimination suggests that it is more likely driven by information than animus.
ASJC Scopus subject areas